R&D Works Up the Options

Aug. 1, 1983

From the local avionics shop to stations in space, technological implications for the future.

The air Force’s technology needs are as immediate and everyday as being able to tell, surely and quickly, what’s wrong when a warning light says there’s an avionics problem in the south end of an airplane.

Other needs are as distant and undefined as preparing man to function in space, perhaps performing on-orbit repair and modification of satellites.

Still another set of requirements suggests the concept of the “autonomous airplane,” in which a pilot might go to war with integrated systems aboard his aircraft feeding him almost any kind of data he can use, helping him make decisions, and freeing him of dependence on links to the ground or to other aircraft.

Even as Air Force Systems Command grapples with its priority objective of cost control (seep.45), it dares not slacken the pace on R&D. In some cases, technological advances may make systems more of affordable. In other cases, they won’t.

“One thing for sure,” says Gen. Robert T. marsh, AFSC Commander, “”if we do not pursue the technology now, we won’t have to worry about the cost/capability tradeoff questions later.”

And the technology must be pursued, because the potential opposition is getting better. A decade ago the United States was ten to twelve years ahead of the Soviet Union in microelectronics and computers. That lead is only three to five years now. No longer does the United States have a monopoly on such capabilities as lookdown/shootdown radar.

The Shelf Gets Bare

A problem here is that the United State has shorted itself on scientific research for many years, and the technology based has eroded badly. “Our society tends to treat technology base issues – developing scientific knowledge and technology – as luxuries or surcharges on basic business,” General Marsh says. “Ethereal luxuries disconnected with today’s reality. That attitude is terribly wrong.”

Continued neglect of the technology base threatens to eat away what remains of the US lead in some critical capabilities.

“Closer to home, one finds that the Air Force is losing its traditional role of technological leadership among the services,” General Marsh said in a state of the command assessment earlier this year. “In recent years the Air Force has lagged behind all the services and DARPA [Defense Advanced Research Projects Agency] in spending for basic and applied research.” Among the reasons why this is disturbing, he says, is that “the most technically advanced service cannot afford to mortgage its future through inadequate attention to the maintenance of a viable weapons technology base.”

With the shelf thus getting bare, General Marsh says, “there are a lot more technologies out there to purse than we in the military can possibly afford to push.” AFSC’s strategy is to go after those technologies that offer the highest payoff with low or limited risk in order to field capabilities the Air Force needs at a cost the nation can afford.

Possibilities in Space

A fair number of the most promising R&D options concerns military applications in space. In the Air Force’s view, space is a place, not a mission, and it may be both feasible and necessary to conduct a broad range of operations from there in the future.

“There are all kinds of space capabilities one can conjure up beyond the limited spacecraft of today that simply go around the earth due to the laws of physics,” General Marsh says.

There is substantial interest in defensive measures, techniques that might allow a satellite to jump out of the way or otherwise defend itself. This interest is driven in part by the fact that the only operational anti-satellite capability in existence belongs to the Soviet Union.

“Survivability is one area that deserves critical attention.” General Marsh says. “We can’t continue indefinitely t expand our reliance on space-based capabilities unless we address the question of their survivability. There are a lot of techniques that bear directly on space survivability that need to be pursued and aren’t being pursued very aggressively today.”

Elsewhere, the prospect of materials with greater heat resistance points to more design flexibility in reentry vehicles. Composites of the future, stronger and lighter than materials of today, could enable the construction of space stations and large platforms with increased sensor or communications capability. Developments in efficient rocket propulsion may permit the repositioning of space systems, including changes in altitude or orbital plane.

“Surely,” General Marsh says, “moving one satellite to another area where it is suddenly needed must be cheaper than building and launching two satellites to achieve the same effect.”

Military Man in Space

The role military man will play in space is uncertain.

“Some maintain that man is simply excess baggage in an on-orbit system,” General Marsh says. “Others believe manned systems having great potential to increase our capabilities, provide flexibility, and do more with less.

“The fact is that we do not yet have the requisite technology programs under way to determine which view is correct. We simply do not know the military-related mission limitations of man in space, nor do we know enough about additional capabilities that might result from man-machine interfaces aboard space systems.”

In any event, work is needed on life-support technologies. General Marsh cites, for example, the requirement for a lifeboat to rescue stranded astronauts.

Later along, such man-in-space activities as on-orbit repair, construction, reconfiguration, and modification of satellites, perhaps using plug-in modules, may turn out to be advisable for reasons of economy as well as capability.

“If we look at the cold, hard facts that are available to us today, we can’t conclude that there’s a useful undertaking or an immediate opportunity for man – military man – in space,” General Marsh says. “Even so, we need to pursue the technologies related to military man in space so we’ll have the opportunity for it if that proves to be a sensible under taking some time in the future.”

Reliable and Ready

Light-years away from space in terms of public attention is the important but unglamorous R & D work of making systems more reliable and ready to go.

“Combat effectiveness and readiness are not determined solely by weapon system capability.” General Marsh says. “Availability – having a weapon system ready when needed – is as important as the capability itself.”

Real progress is being made. The F-15, for example, can surge to better than four sorties a day, compared with an average of one sortie every four days for World War II fighters. Subsystems are better too. The latest UHF radios have a mean time between failures of about 1,000 hours. UHF radios used to fail within thirty to 100 hours.

“There was the era when we had just a great proliferation of lousy, low-reliability avionics,” General Marsh says. “We generally felt we ought to develop some high-reliability avionics and standardize on them. And we’ve come a long way.

“You’ll see the same UHF radios, ARC-164s, in all of our first-line airplanes. You’ll see that good new TACAN in all of them. In all of those that need LORAN, you’ll se that good LORAN-101, and so on.”

For all of its benefits in reliability, logistics, and cost, standardization alone – or applied with a sledgehammer – is not the answer. The danger is freezing on obsolescence. Unrelenting standardization may block the emergence of newer and better systems.

“You’ve got to be somewhat careful that you don’t let your standardization objective impede progress either toward more capability per unit of cost,” General Marsh says. “Also, we want to keep moving forward. That’s why we’re looking at ring laser gyros instead of the old mechanical type.”

Modular Avionics

They way around this problem is to standardize, not on the avionics black boxes themselves, but on the receptacles they fit into on the airplane, and on the way they hook together. The shorthand for this is “modular avionics.”

“That’s the architecture of our new aircraft,” General March says. “We’re going to have standard multiplex buses and standard interfaces. People can then develop avionics improvements within form, fit, and function and you plug them in and it’s not a major overhaul or modification of the aircraft.”

Standard interfaces for ordnance are coming, too.

“All weapons producers or developers will put the standard interface on the weapon, and we’ll have the standard interface at the pylon or the rack or wherever,” General Marsh says. “There’ll be pins there for signal, pins for power, a fiber optics connection if the weapon needs that – and so on into the aircraft system, hopefully by way of the multiplex bus. Now then, if you develop a new weapon, we know darn well we won’t have to string a bunch of new wiring out through the airplane. The weapon will fit. If you need twenty-eight volts, you go to pin J. If you need 110 volts, 400 cycle, you to go to pin K.

Modular avionics will make modifications cheaper as well as easier. The air Force tends to keep its basic airframes for a long time, but goes through several generations of avionics or them. In the future, the inevitable upgrades can be handled with less ripping and tearing. In addition, standardization will probably mean more competition from industry to make the new module. No longer will unique features of the system lead almost inevitably back to the producer of the previous module.

A Fix for CND/RTOK

This is not to say the Air Force has solved all of its avionics problems.

A particular hair shirt in the maintenance business in “CND/RTOK.” It stands for “Could Not Duplicate/Retest OK,” and it’s what the people in the avionics shop write on the form when they can’t find the problem that the warning sensor said was present when the airplane was in flight.

With the built-in test/fault isolation (BIT/FI) equipment now available, the best the shop may be able to do is narrow the problem down to one of two black boxes. CND/RTOK rates are high, and avionics maintenance expenses are steep. Even worse, the box may be put back into service with the non-improbable reasoning that the fault was in the warning sensor, no in the black box. If the box fails repeatedly, the airplane may be grounded for the costly business of tearing apart one thing after another in search of the problem.

The solution may lie in very-high-speed integrated circuits (VHSIC), which have the capability to carry built-in testing down to the chip level of every black box in the avionics suite.

“We think this will be the enabling technology to allow us to get down to two-level maintenance,” says Maj. Mike Borky of AFSC’s DCS/Science and Technology. “Do away with the intermediate avionics shop. The system boldly announces where is has failed. The crew chief or whoever opens up the cowling and looks for the blinking red light. He pulls that ‘cigarette pack’ out. Depending on how much it cost, he either throws it away or drops it in the Return to Depot bag. He plugs in a new one and closes it up.”

The distinction between line-replaceable units and shop-replaceable units may disappear, and in the time, manpower-intensive avionics maintenance may fade as the driving factor in aircraft support requirements.

Self-healing Systems

Computational technology is also leading toward the day of fault-tolerant electronics, which, in an over-simplified sort of way, might be thought of as self-healing systems.

If an avionics system fails in flight, the computer would reprogram itself, usually diverting to redundant capacity there for that express purpose. “You may not even know while you’re in the heat of battle that something has failed,” Major Borky says.

If failures are so extensive that they swamp the regular backup circuits, the machine would alert the pilot to the problem and flash possible work-around solutions on his head-up display. If he must continue using systems A and B, then he has to turn off either C or D because there isn’t enough capacity left to support them all.

These concepts call for near-total integration of avionics, and the Air Force’s existing multiplex bus standard, MIL STD 1553B, is probably not adequate for the fully integrated airplane. Preliminary analysis indicates that significantly higher data rates will be needed. The 1553B architecture may be used for subsystems that can live with the lower data rate, and higher speed mux buses may be added for those that can’t.

The VHSIC Cornerstone

The cornerstone for many of these plans – and for much else the Air Force wants to do – is VHSIC. A 1981 Defense Science Board panel, looking at what if called “order of Magnitude” technologies, assigned its top figure of merit (based on the ration of high payoff to low development risk) to VHSIC.

VHSIC was originally billed as the next generation of speed in computer chips, but that doesn’t say it all.

“The emphasis is on high throughput of data,” says Major Borky. “The functional throughput rate is the product of speed and chip density. You can get throughput both ways. As you make transistors smaller, they switch faster. It’s a fundamental law of physics, and it means that as chips get denser, they also get faster.”

The first few VHSIC chips are just now rolling off the line, but dozens of studies are already identifying benefits to systems that include size, power consumption, and reliability as well as enhanced performance.

“VHSIC lends itself to doing some very special military tasks,” General Marsh says. “We can’t do without something like it. For example, if you want to make darn smart terminal seekers today, you’re limited in a lot of munitions for volume and weight. VHSIC starts to open up putting the real smarts in terminal guidance systems that you can’t really do today simply because of weight and power and space limitations.”

He says that getting signal processing capabilities into smaller packages with VHSIC will be a considerable benefit. Better IFF (Identifications, Friend or Foe) also looms as a possibility.

“Notionally, if you could integrate all of the characteristics of an enemy target, perhaps its signal or heat characteristics, even down to making a shape identification, you can characterize that target very well,” General Marsh says. “Then if you have the processing capability to put all of that together, I think if will make identification possible. VHSIC will open up the processing power to facilitate enemy identification.”

An even greater advantage may come in tactical data fusion. C3I planners have long talked of was to handle the vast amounts of information a modern battlefield generates, but the real answer has kept eluding them.

“The challenge is still to get that information back from all those sensors, chew on it fast and draw out from it the essential elements, and then provide it to a decision-maker so he can act on it quickly,” General Marsh says. “VHSIC is going to give us the power to do that. And not necessarily in a big ground station that can be done away with by one little 500-pound bomb. It might be done in a C3 airplane, or it’s thinkable that it could be done in a single fighter or a bomber.”

Artificial Intelligence

Even more amazing capabilities might be packed into a single airplane as computer hardware advances and as a radically new software technique – artificial intelligence – matures. Artificial intelligence is the computer simulation of thought or decision-making processes that normally require human intelligence. The computer adapts its own programming as it “learns” more about its environment.

“Previously,” says Dr. Bernard Kulp, AFSC Chief Scientist, “Computers manipulated numbers. They may be very clever manipulators of numbers, but, by and large, they just add and subtract. Now when one talks about artificial intelligence techniques, one doesn’t store numbers in a computer or do mathematical manipulations. One stores knowledge – bits and pieces of things that basically represent knowledge – in the computer.”

The computer takes statements of knowledge, compares them, and infers directly a further statement of knowledge. “There is a knowledge train.” Dr. Kulp says. “If this happens, and this and this, then this ought to happen. If it doesn’t, try this and this and this.”

Significant Air Force involvement with artificial intelligence began about three years ago. Prior to that, DARPA was the main defense participant in this line of research.

“Artificial intelligence is still a science,” Dr. Kulp says. “I’m not ready to call it a technology yet. Technology to me is the state of affairs when one is ready to apply it and do something with it.”

The Air Force’s thrust n artificial intelligence is not the same as that of the general scientific community. “The Air Force’s job now is reduction to practice, with limited objectives,” Dr. Kulp says. “I think the scientific community is not interested particularly in reduction to practice, except as a hobby. Their prime interest is in advancement of the state of the art.”

Among the likely candidates for applying artificial intelligence to Air Force problems are information fusion, very reliable avionics, diagnosis of system faults, and even a “software production assistant” to help humans with their computer programming. The most ambitious and fascinating application being considered is the “autonomous airplane.”

The Autonomous Airplane

It begins with the integration of all the individual sensors and pieces of avionics equipment, each of which has been becoming progressively more capable in its own right. Their synergistic capability is greater still. There is an interchange of information generated here and there all over the sensor suite. Every part of the system can take advantage of computational power available in any part of it.

“With these kinds of capabilities, we begin to think of doing a lot more things on board the aircraft, things we have traditionally done on the ground and transmitted into the aircraft,” Dr. Kulp says. “One begins to think of the autonomous airplane.”

The concept, he emphasizes, is not one of robotics. There will be a pilot in the cockpit with a new super-smart machine to help him. “Autonomous” refers to the extraordinary extent to which all necessary capabilities can b contained on board and the extent to which the pilot is freed from reliance on sources external to his aircraft. Artificial intelligence may be much smarter than its number-crunching ancestors, but Dr. Kulp believes it will be long, long time before an unmanned system that will consistently defeat a man will be built.

“The important thing we need to do with this autonomous aircraft concept is work on what to automate,” he says. “We know how to automate almost everything. We just don’t know yet what to automate so that the man-machine combination is indeed optimized for capability in any situation.”

Autonomous navigation, threat analysis, and target recognition would stand high on the list of candidates. Because of its artificial intelligence features, the machine would give the pilot information instead of raw data. Much of that information would probably come in the form of situation reporting, presentation of options, and probabilities connected with various courses of action.

“We see on-board sensors as having automatic target recognition to the extent, first, that it can tell tank from a truck and know it’s not a boulder or a puddle of water,” Dr. Kulp says. “Eventually, as we move down the line twenty or thirty years, identification of a particular model number may be possible.”

Autonomous navigation systems might be on the lookout for a particular bridge, knowing that the one it wants has four spans, is of open truss construction, and has railroad tracks but no highway. The machine takes it all in, keeps subtotaling and adding to what it knows. What the navigation subsystem knows, the threat assessment subsystem knows, too. If part of the bridge is blown away in battle, the artificial intelligence system sees, learns, and remembers. It might conclude that the damage has put the railroad out of commission, so the equipment it saw loaded on the train east of the bridge will be late in getting to its destination west of the bridge.

Engines and Airframes

By the time the autonomous airplane is established in the fleet, considerable strides will have been made in engines and airframes.

“We see a twelve-to-one thrust-to-weight ration coming out in twenty years if we pay attention to our business,” Dr. Kulp says. “We see aerodynamic drag coming down. We can expect to see the drag at Mach 1.5 to Mach 2 being not much higher than it is at Mach 0.9 right now. That, combined with the trust-to-weight ratio of the engines and some improvement in specific fuel consumption, says that sustained operation at supersonic speeds will be economically achievable in military aircraft.”

Continued progress in composite materials technology will improve the strength-to-weight ratio of airframes and allow the tailoring of aerodynamic surfaces in manufacturing.

“We will be able to make very complex surfaces economically, because basically, you make a mold for these things and cast them,” Dr. Kulp says. “It isn’t like you had to machine every square inch of it like we presently do with metal structures. It also allows us to tailor the strength in the high stress direction.”

These features, however, lie beyond the turn of the century. The advanced technology fighter for the 1990s will have improvements on a more modest scale. The engine, for example, may very well have fewer parts than present engines and may be significantly more durable.

“Durability is number one on our list,” General Marsh says. “We’ve jut got to get the operation and support costs of our engines down, because they’ve got some very big logistics tails today. There’s no question but that we want to get the trust-to-weight ratio of our engines up, so lighter-weight materials, in conjunction with durability, are certainly an important objective. We’re looking for good fuel economy, because fuel’s becoming a bigger and bigger factor in the O&S costs. And I would also say good, reliable accessories of all kinds, including fuel control systems.”

He says the Air Force is still looking at the need for supersonic cruise in the advance fighter, and may want the engine to be able to support that.

“And we want thrust,” he says, “but we’re not looking for a big leap forward in thrust from the current fighter engine.”

More R&D of Significance

Among the many other Systems Command R&D enterprises, the following stand out as particularly significant.

  • Stealth, or “low observables,” technology to make aircraft more difficult for the enemy to detect.
  • Autonomous guided weapons. Tactical missiles that lock on after launch with no data links required.
  • Voice controls and unconventional flight paths for fighter aircraft, previewed in the exploits of the AFTI/F-16 (see “The Future Forms Up at ASD,” Air Force Magazine, January ’83).
  • Generators and accelerators for directed-energy weapons.
  • A new guidance system for the small, single-warhead ICBM. It will have to be smaller and lighter than the MX guidance package, and, thanks to new technology such as ring laser gyros, perhaps more accurate as well.